Macroeconomics is divided into (short run) business cycle theory and (long run) growth theory.
Those of us who do business cycle theory have a bit of an inferiority complex (though you might not know it from listening to us argue). Because growth theory seems to be so much more important. Where would you rather live: in a rich country during a recession; or in a poor country during a boom? (Watch the flows of people voting or attempting to vote with their feet if you are not sure how most people would answer.) In the long run, productivity is about the only thing that matters.
We would feel better about ourselves, and what we are studying and teaching, if we could argue that taming the business cycle would improve long run growth.
Notice that I have deliberately personalised this question to make you aware of my personal bias. Macroeconomists like me, who do short run business cycle theory, want to think that what we are doing is important. We want to argue that taming the business cycle would improve long run growth.
(The Great Recession was great for my sort of macro; we haven't had so much fun since the 1970's. The Great Moderation was a boring time for macroeconomists like me, when we seemed to be victims of our own success; all the growth theorists were stealing our limelight.)
Why might business cycles lower the long run growth rate?
Investment falls in a recession, and it is easy to understand why that might happen. If it's hard to sell the goods you are currently producing, why invest to be able to produce even more goods?
The same would presumably be true for investment in human capital. Young workers find it harder to get jobs as they enter the labour market, and so don't get the on-the-job training. And human capital depreciates if it's not used, because we forget stuff, and don't keep up with the new stuff. Though the opportunity cost of going to school is lower if you can't get a job, which will have an offsetting effect. (And by the way, this is one more illustration of the inherently monetary nature of recessions, because a student who "buys" his own time to produce human capital is engaged in home production of investment goods, where no money changes hands; it's very different from a firm that buys investment goods from other firms.)
So we can understand why a recession might have long-lasting or even permanent effects on the level of output. We call that "hysterisis".
Unfortunately for us short run macroeconomists, hysterisis isn't enough.
Investment is low in a recession, but it is also high in a boom. Maybe it all just cancels out, so the average level of investment would be exactly the same if we tamed the business cycle?
If there's diminishing marginal utility of consumption, and increasing marginal disutility of employment, taming the business cycle fluctuations in output and employment would be a Good Thing, but if it doesn't affect the long run growth rate it still doesn't matter much. People would still rather live in a rich country with business cycles than a poor country without.
What we need is some sort of asymmetry between recessions and booms. Better yet, a model with recessions but no booms, so that recessions are valleys in a plateau, and if we could eliminate the business cycle then we could live forever on Irving Fisher's "permanent plateau". We need something like Milton Friedman's "plucking model", which is very different from the symmetry of the standard New Keynesian model. And we need to argue that investment too would be higher on a permanent plateau.
It can be done; my colleague Vivek Dehejia and I built such a model once. It differed from the supply-side of the standard New Keynesian set-up in only one relevant respect: monopolistically competitive firms had to produce output before observing the aggregate demand shock. So if the demand shock was positive, and large enough, it was too late for firms to expand output to meet demand, and there would be queues of customers unable to buy goods. (Prices were set in advance too, of course.) This meant that booms were smaller than recessions, so average output would be higher if the business cycle were tamed. And more importantly, the expected marginal revenue product of capital would be higher if the business cycle were tamed, so investment would be higher, and (in an AK model) the growth rate would be permanently higher too (and so would the rate of interest).
OK, I confess that one of the purposes of this post is to plug my old paper with Vivek. But there's a more general point to be made here, that goes beyond the rather clunky assumption we made in our model.
When the business people who make investment decisions say they don't like "uncertainty", maybe they are trying to tell us something about the effect of business cycles on investment, and saying something more than just "we invest less in a recession and more in a boom". And if us short run macroeconomists want to argue that we are important for long run growth too, maybe we should be thinking about the various different ways this might work.